Computational methods are increasingly being recognized as valuable tools for the study of biomolecular structure and function. Advances in simulation techniques, most notably in the areas of conformational sampling, fast electrostatics, molecular dynamics integration, and quantum-mechanical calculations, are having significant impact on structural biology. New algorithmic approaches, hierarchical spatial representations, and improved computing platforms will continue to enhance the reliability of macromolecular simulations and to increase their applicability and relevance to biomedical research. The purpose of our multidisciplinary workshop on Methods for Macromolecular Modeling M3--the fourth in a series of international meetings (1994: Lawrence, Kansas; 1997: Berlin, Germany; 2000: New York, NY)--is to bring together both developers of computational tools for biomolecular simulations and those biological and chemical scientists who utilize (or are interested in applying) computer modeling to macromolecular problems. The workshop will be held at the University of Leicester on August 18-21, 2004 and is co-sponsored by SIAM (Society for Industrial and Applied Mathematics), as part of SIAM's new Activity Group in the Life Sciences. As evidenced by the success of the prior meetings--in terms of participant number, tangible outcomes, influence on young researchers, and reviews of our collected volumes--our workshop has made important impact on the community in both research advances and education. These conference goals and themes resonate well with many activities at the NIGMS Center for Bioinformatics and Computational Biology as well as several additional NIH roadmap themes/groups, including structural biology and interdisciplinary research (innovations in technology and methods; removing structural barriers). The topics to be highlighted in 2004 are: (1) Algorithms for force evaluation, integration, optimization, and sampling; (2) Methods for structure prediction, reaction paths, free energy profiles, and conformational dynamics; (3) Representation of force fields and implicit solvent; (4) Multiscale techniques for quantum/classical and classical-elastic models; and (5) Applications to enzyme catalysis, DNA modeling, and DNA/protein systems. The program will provide a timely and unique opportunity for close interaction and scientific exchange among biomolecular researchers, computer scientists, and applied mathematicians. In our genomic revolution era and the integrative sciences it has propelled (e.g., proteomics, cellomics and networks; metabolomics, and biotechnology), such synergy will kindle new ideas and help prepare young scientists for crossdisciplinary research at the interface of computational science and biology. The program will emphasize the application of computational methods to problems of medical relevance, such as the consequences of protein folding advances and DNA polymerase fidelity to disease mechanisms and new therapeutics. The assessment of current progress and the identification of future directions in the field will be accomplished through a first day of "Focus Sessions," including key presentations and panel discussions, from which a report will be prepared. As in the prior meeting, we expect to collect articles by invited speakers for publication in Springer Verlag's Lecture Notes series in Computational Science & Engineering (LNCSE); the volume will be edited by selected members of the organizing committee. These tangible records, in addition to a carefully designed program, are expected to serve the community and the funding agencies by educating junior scientists at interdisciplinary interfaces, stimulating new ideas in computational techniques for biomedical modeling, and identifying key areas for future research. In addition, recruitment of speakers and supported travelers will be guided by the goal of obtaining a good representation for underrepresented groups (women, minorities, and scientists with disabilities) by proactively encouraging such participants.